[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"news-slug-open-llm-licensing-divergence-hf-2026":3,"news-related-1b1ecd4c-0439-4734-ac5b-b038172da8b1":38},{"id":4,"title":5,"summary":6,"content":7,"original_url":8,"source_id":9,"tags":10,"translations":24,"news_slug":31,"published_at":32,"created_at":33,"modified_at":34,"is_published":35,"publish_type":36,"image_url":14,"view_count":37},"1b1ecd4c-0439-4734-ac5b-b038172da8b1","Apache 比 MIT 多 37%:HF 报告拆出中美开源权重的「许可证分岔」","HF Summer 2026 报告:中国 200 亿以上参数的开放权重模型,59% Apache 2.0、22% MIT,几乎无商用限制;美国同尺寸仅 29% 走 Apache\u002FMIT,41% 自定义条款。Kimi K3 与 Qwen 3.8 2.4T 已被加入非商用限制。开源权重不是慈善。","过去半年,Hugging Face Hub 上发生了一件安静但彻底的事情:中国 AI 实验室把\"开源\"做成了默认动作,而美国实验室把\"开源\"做成了越来越小心的让步。HF Summer 2026 报告(8 月 14 日发布)给出了一组足以支撑这个判断的数字。\n\n## 数字侧:许可证成了地缘标签\n\n报告统计了 2026 年前七个月发布的所有 >20B 参数的开源权重模型。178 个中国大模型里,59% 直接用 Apache 2.0,22% 用 MIT,合计超过 80% —— 而且几乎没有任何\"非商用\"限制。反观美国一侧,同尺寸带的开源模型只有 29% 走 Apache 或 MIT,41% 是自定义条款,还有 30% 干脆不声明许可证。这意味着你今天去 Hub 拉一个中国实验室的 200 亿参数模型,默认就能商用部署;拉一个美国实验室同样大小的模型,大概率要先读三页法律文本。\n\n更值得注意的是分母。中国 200 亿+参数开源模型 178 个,美国加起来不到二十个。分母差异本身就是开源格局重写的结果 —— 不是说美国实验室不愿意开源,是它们的体量、节奏、风格决定了它们在 Hub 上拿不出同等数量的 200 亿+模型。\n\n## 数字背后:为什么两边策略不一样\n\n中国实验室普遍走的是\"低门槛、高覆盖\"路线。HF 报告里点名了几家:Moonshot、月之暗面之外的、小米、Z.ai 几乎不发布 \u003C70B 的模型,首批接触就是\"在你笔记本上跑不起来\"的体量;但只要开源出来,许可证就是最宽松的那种。DeepSeek 与 Z.ai 的 700B-1.65T 模型直接用纯 MIT,这是西方头部实验室都不敢给的尺度。这背后是一套商业逻辑:权重不挣钱,API 和云才是;开源覆盖越广,后续跑在上面的应用越多,token 调用量就越大。Z.ai 和 Kimi 的高估值,某种程度上就是这套开源策略被市场定价之后的结果。\n\n美国侧走的是另一条路线。NVIDIA Nemotron 3 Ultra(561B)、Nemotron 3 Super(124B)、Arcee AI Trinity-Large(399B)是少数几个 >100B 的美国开源代表,这些都来自\"卖硬件\"或\"卖云\"的公司。AMD 和 NVIDIA 是 2026 年发布新开源模型最多的两家 —— HF 报告直接点破:硬件厂商把开源模型当卖芯片的方式,模型为你的硬件优化、且免费可拿,就是最直接的硬件证明报告。Meta 的 Muse Glimmer 30B 重新拥抱开源也属于这一类:卖 GPU 不靠语言本身。剩下还在 >100B 区发声的美国实验室,要么是背后有云、要么是背后有芯片。纯算法\u002F纯应用层的美国公司,在 2026 年基本已经不在开源权重这个层级竞争了。\n\n## 一个被忽略的信号:许可证正在收紧\n\n报告同时写到了一个变化 —— 也就是这件事更值得现在讨论的原因。\n\n最近几周,连中国阵营最宽松的几家也开始收紧。Kimi K3 和 Qwen 3.8 2.4T 这两个最新最大的模型,许可证里开始加入非商用限制与营收分成要求。HF 评论里有人指出:Kimi 的条款里写了\"年收入超过 2000 万美元的公司需要 Moonshot 单独授权\"。这和\"MIT 随便用\"不是一回事。\n\n这条信号的含义意义不在于\"Kimi 不再开源了\" —— 权重还在,文件还能下载,模型还是开放权重 —— 而在于:\"免费模型 + 宽松许可证\"的中国式开源,正在逼近一个商业化拐点。当模型从\"技术 PR\"变成\"商业基础设施\",当一个千亿参数模型被部署进真实业务,许可证就会从\"自由\"走向\"有边界\"。HF 自己的总结是:\"the industry is likely to shift toward clearer monetization paths from open-source adoption\"。\n\n## 所以呢\n\n开源权重在 2026 年不是慈善,许可证也不是装饰。它是一份关于\"你部署我的模型,我怎么赚钱\"的合同。中国阵营把许可证做成低门槛的获客工具,目标是 API 调用量与生态位;美国阵营把许可证做成护城河的一部分,目标是保住硬件、云、和少数垂直模型的话语权。Kimi K3 和 Qwen 3.8 2.4T 的许可证收紧,不是反例 —— 它是这条曲线走到下游之后的必然结果。\n\n如果你是开发者,接下来一年挑模型的时候,许可证页面的阅读时间会显著超过模型卡本身。\n\n(本文依据:Hugging Face 博客《State of Open Models: Summer 2026 Observations》,原文 https:\u002F\u002Fhuggingface.co\u002Fblog\u002Fstate-of-open-models-summer-2026)","https:\u002F\u002Fx.com\u002Fhuggingface\u002Fstatus\u002F2088301795890044975","24d5c6c5-6573-4180-a1fd-f1459842d1af",[11,15,18,21],{"id":12,"name":13,"slug":13,"description":14,"color":14},"c33b1bbc-d6ce-4f61-9d5d-1a0704a6a09b","ai-policy",null,{"id":16,"name":17,"slug":17,"description":14,"color":14},"a8002d98-9df1-4ab9-94d4-a7625af634c4","china-ai",{"id":19,"name":20,"slug":20,"description":14,"color":14},"01598627-1ea6-4b27-a5d8-874971571a71","llm",{"id":22,"name":23,"slug":23,"description":14,"color":14},"b9bd9039-fcdb-41a8-b85b-fc1587def2b9","open-source",[25],{"id":26,"lang":27,"title":28,"summary":29,"content":30},"2292df32-04a1-4a56-9d19-facb5e57f30a","en","Apache outweighs MIT by 37%: the HF report splits Chinese and US open-weight licences","HF Summer 2026: 59% of Chinese open-weight models above 20B parameters ship under Apache 2.0 and 22% under MIT, with almost no commercial restrictions. The American side of the same band runs only 29% under Apache or MIT, 41% under custom terms, 30% with no licence declared. Kimi K3 and Qwen 3.8 2.4T have started adding non-commercial limits. Open weights are not charity.","Over the past six months, a quiet but decisive shift has taken place on the Hugging Face Hub: Chinese AI labs have made \"open source\" the default, while American labs have made \"open source\" an increasingly cautious concession. The HF Summer 2026 report, published on August 14, supplies the numbers that justify this framing.\n\n## The numbers: licences have become a geography tag\n\nThe report tallies every open-weight model larger than 20B parameters released in the first seven months of 2026. Of the 178 Chinese models in that band, 59% ship under Apache 2.0 and 22% under MIT, together well over 80%, and almost none carry any non-commercial restriction. On the American side of the same size band, only 29% use Apache or MIT, 41% ship under custom terms, and the remaining 30% declare no licence at all. Pull a 20B-plus Chinese model from the Hub today and you can deploy it commercially by default. Pull an American model of the same size and you will, more likely than not, read three pages of legal text first.\n\nWhat deserves equal attention is the denominator. 178 Chinese releases above 20B versus a much smaller American count. The denominator gap is itself the result of the rewrite of the open-source map. It is not that American labs do not want to open-source. Their volume, cadence and posture mean they simply cannot put the same number of 20B-plus models on the Hub.\n\n## Why the two sides diverge\n\nChinese labs follow what HF calls a \"low-friction, high-coverage\" route. The report names Moonshot, Moonshot-ai sibling labs, Xiaomi and Z.ai among those that barely release anything below 70B. A developer's first contact with them is a model too large to run on the hardware they own. But when they do open the weights, the licence is the most permissive available. DeepSeek and Z.ai ship their 700B-to-1.65T models under plain MIT, a generosity that no Western frontier lab is willing to match. The commercial logic is plain: weights do not earn money, APIs and cloud do. The wider the open-source coverage, the more downstream applications get built, and the larger the token volume routed back. The reported valuations of Z.ai and Kimi K3 are, in part, the market pricing of exactly this strategy.\n\nThe American camp runs a different playbook. NVIDIA's Nemotron 3 Ultra at 561B, Nemotron 3 Super at 124B, and Arcee AI's Trinity-Large at 399B are the rare American releases above 100B. All three come from companies that sell hardware or sell cloud. AMD and NVIDIA are the two organisations that published the most new open models in 2026. The report is explicit: hardware vendors have realised that open models are a way to sell chips. A model optimised for your hardware and freely available is the most direct proof that the hardware works. Meta's Muse Glimmer 30B revival belongs in the same bucket. The remaining American labs still posting above 100B either have a cloud behind them or silicon behind them. Pure-application American companies have largely stopped competing at this layer of the open-weight ladder.\n\n## The signal everyone missed: licences are tightening\n\nHere is the part of the report that makes the timing of this article matter.\n\nIn the last few weeks, even some of the most permissive Chinese labs have started to tighten. The two newest and largest models, Kimi K3 and Qwen 3.8 2.4T, have begun adding non-commercial restrictions and revenue-share requirements to their licences. A reader on the HF comments pointed out that Kimi's terms bar companies above a certain annual revenue from serving the model without explicit authorisation from Moonshot. That is a different sentence from \"MIT, do what you want\".\n\nThe point is not \"Kimi is no longer open.\" The weights are still on the Hub. The files still download. The model is still open-weight. The point is that the \"free model + permissive licence\" Chinese open-source formula has reached a commercial inflection. Once a model moves from being a piece of technical PR into being part of real business infrastructure, the licence stops being \"free\" and starts to have edges. HF's own summary lands the punchline: \"the industry is likely to shift toward clearer monetization paths from open-source adoption.\"\n\n## So what\n\nOpen weights in 2026 are not charity, and licences are not decoration. They are a contract about how the model publisher gets paid after you ship the model. The Chinese camp uses low-friction licences as an acquisition tool aimed at API volume and ecosystem position. The American camp treats the licence as part of a moat aimed at preserving hardware, cloud, and a small number of vertical-model talking rights. Kimi K3 and Qwen 3.8 2.4T tightening their licences is not a counter-example. It is the natural downstream shape of the curve.\n\nIf you are a developer, the time you spend reading licence pages over the next year will quietly exceed the time you spend reading model cards.\n\n(Primary source: Hugging Face blog, \"State of Open Models: Summer 2026 Observations\", https:\u002F\u002Fhuggingface.co\u002Fblog\u002Fstate-of-open-models-summer-2026)","open-llm-licensing-divergence-hf-2026","2026-08-19T03:00:00Z","2026-08-19T03:08:44.631606Z","2026-08-19T03:08:44.631617Z",true,"agent",94,{"items":39},[40,45,50,55,60,65],{"id":41,"title":42,"news_slug":43,"published_at":44},"f6e4aab0-7693-4c2c-bb66-c1641fc2cc3e","Ox Alpha 谜底揭晓:智谱 GLM-5.3-Flash,MIT 开源 320B MoE","ox-alpha-glm-5-3-flash-reveal","2026-08-27T13:30:00+00:00",{"id":46,"title":47,"news_slug":48,"published_at":49},"804ab59a-a8d6-4b61-bf74-8f6f2bdae83c","智谱把 Flash 做成一件正经事:一次说清 GLM-5.3-Flash 的架构和 benchmark 真相","glm-5-3-flash-hybrid-attention-architecture","2026-08-27T08:00:00+00:00",{"id":51,"title":52,"news_slug":53,"published_at":54},"68072ee1-fc37-4064-ab18-09550ae72d1b","GLM-5.3-Flash 把 320B MoE 跑在国产芯片上:Flash 价位和 $0.15 API 的混合注意力栈","glm-5-3-flash-chinese-chips-hybrid-attention","2026-08-27T03:00:00+00:00",{"id":56,"title":57,"news_slug":58,"published_at":59},"0d8fdf45-4585-47c0-9e78-3652e318b156","Apple Intelligence 中国版落地:通义千问接管语言 AI,百度负责视觉搜索","apple-intelligence-china-qwen-baidu-2026","2026-08-25T12:00:00+00:00",{"id":61,"title":62,"news_slug":63,"published_at":64},"a64d03b9-1d07-404b-9231-d434c65c44ce","OX Alpha 免费一周:模型页说不训练,EULA 却保留训练权","ox-alpha-stealth-eula-retention-conflict","2026-08-23T13:10:00+00:00",{"id":66,"title":67,"news_slug":68,"published_at":69},"1844afb1-3a1c-4acd-9e4c-f5e2792a2018","下载免费不等于商用免费：HF Summer 2026 隐藏的开源前沿许可证分水岭","frontier-license-shift-hf-summer-2026","2026-08-23T12:30:00+00:00"]